Probabilistic graphical models (PGMs) are a rich framework for encoding probability distributions over complex domains: joint (multivariate) distributions over large numbers of random variables that interact with each other. These representations sit at the intersection of statistics and computer science, relying on concepts from probability theory, graph algorithms, machine learning, and more. They are the basis for the state-of-the-art methods in a wide variety of applications, such as medical diagnosis, image understanding, speech recognition, natural language processing, and many, many more. They are also a foundational tool in formulating many machine learning problems.
This course is the third in a sequence of three. Following the first course, which focused on representation, and the second, which focused on inference, this course addresses the question of learning: how a PGM can be learned from a data set of examples. The course discusses the key problems of parameter estimation in both directed and undirected models, as well as the structure learning task for directed models. The (highly recommended) honors track contains two hands-on programming assignments, in which key routines of two commonly used learning algorithms are implemented and applied to a real-world problem.
Этот курс входит в специализацию ''Специализация Графические вероятностные модели '
Probabilistic Graphical Models 3: Learning
от партнера
Об этом курсе
4.6
Оценки: 196
Приобретаемые навыки
AlgorithmsExpectation–Maximization (EM) AlgorithmGraphical ModelMarkov Random Field
Программа курса: что вы изучите
Неделя
116 минуты на завершение
Learning: Overview
This module presents some of the learning tasks for probabilistic graphical models that we will tackle in this course.
1 видео ((всего 16 мин.))
1 видео
Learning: Overview15мин
1 ч. на завершение
Review of Machine Learning Concepts from Prof. Andrew Ng's Machine Learning Class (Optional)
This module contains some basic concepts from the general framework of machine learning, taken from Professor Andrew Ng's Stanford class offered on Coursera. Many of these concepts are highly relevant to the problems we'll tackle in this course.
6 видео ((всего 59 мин.))
6 видео
Regularization: Cost Function 10мин
Evaluating a Hypothesis 7мин
Model Selection and Train Validation Test Sets 12мин
Diagnosing Bias vs Variance 7мин
Regularization and Bias Variance11мин
2 ч. на завершение
Parameter Estimation in Bayesian Networks
This module discusses the simples and most basic of the learning problems in probabilistic graphical models: that of parameter estimation in a Bayesian network. We discuss maximum likelihood estimation, and the issues with it. We then discuss Bayesian estimation and how it can ameliorate these problems.
5 видео ((всего 77 мин.)), 2 тестов
5 видео
Maximum Likelihood Estimation for Bayesian Networks15мин
Bayesian Estimation15мин
Bayesian Prediction13мин
Bayesian Estimation for Bayesian Networks17мин
2 практического упражнения
Learning in Parametric Models18мин
Bayesian Priors for BNs8мин
Неделя
221 ч. на завершение
Learning Undirected Models
In this module, we discuss the parameter estimation problem for Markov networks - undirected graphical models. This task is considerably more complex, both conceptually and computationally, than parameter estimation for Bayesian networks, due to the issues presented by the global partition function.
3 видео ((всего 52 мин.)), 2 тестов
3 видео
Maximum Likelihood for Conditional Random Fields13мин
MAP Estimation for MRFs and CRFs9мин
1 практическое упражнение
Parameter Estimation in MNs6мин
Неделя
317 ч. на завершение
Learning BN Structure
This module discusses the problem of learning the structure of Bayesian networks. We first discuss how this problem can be formulated as an optimization problem over a space of graph structures, and what are good ways to score different structures so as to trade off fit to data and model complexity. We then talk about how the optimization problem can be solved: exactly in a few cases, approximately in most others.
7 видео ((всего 106 мин.)), 3 тестов
7 видео
Likelihood Scores16мин
BIC and Asymptotic Consistency11мин
Bayesian Scores20мин
Learning Tree Structured Networks12мин
Learning General Graphs: Heuristic Search23мин
Learning General Graphs: Search and Decomposability15мин
2 практического упражнения
Structure Scores10мин
Tree Learning and Hill Climbing8мин
Неделя
422 ч. на завершение
Learning BNs with Incomplete Data
In this module, we discuss the problem of learning models in cases where some of the variables in some of the data cases are not fully observed. We discuss why this situation is considerably more complex than the fully observable case. We then present the Expectation Maximization (EM) algorithm, which is used in a wide variety of problems.
5 видео ((всего 83 мин.)), 3 тестов
5 видео
Expectation Maximization - Intro16мин
Analysis of EM Algorithm11мин
EM in Practice11мин
Latent Variables22мин
2 практического упражнения
Learning with Incomplete Data8мин
Expectation Maximization14мин
4.6
Рецензии: 3143%
начал новую карьеру, пройдя эти курсы
33%
получил значимые преимущества в карьере благодаря этому курсу
20%
стал больше зарабатывать или получил повышение
Лучшие рецензии
very good course for PGM learning and concept for machine learning programming. Just some description for quiz of final exam is somehow unclear, which lead to a little bit confusing.
Great course! Very informative course videos and challenging yet rewarding programming assignments. Hope that the mentors can be more helpful in timely responding for questions.
Преподаватели
Daphne Koller
ProfessorSchool of Engineering
О Стэнфордский университет
The Leland Stanford Junior University, commonly referred to as Stanford University or Stanford, is an American private research university located in Stanford, California on an 8,180-acre (3,310 ha) campus near Palo Alto, California, United States.
О специализации ''Графические вероятностные модели '
Probabilistic graphical models (PGMs) are a rich framework for encoding probability distributions over complex domains: joint (multivariate) distributions over large numbers of random variables that interact with each other. These representations sit at the intersection of statistics and computer science, relying on concepts from probability theory, graph algorithms, machine learning, and more. They are the basis for the state-of-the-art methods in a wide variety of applications, such as medical diagnosis, image understanding, speech recognition, natural language processing, and many, many more. They are also a foundational tool in formulating many machine learning problems.
Часто задаваемые вопросы
Когда я получу доступ к лекциям и заданиям?
Зарегистрировавшись на сертификацию, вы получите доступ ко всем видео, тестам и заданиям по программированию (если они предусмотрены). Задания по взаимной оценке сокурсниками можно сдавать и проверять только после начала сессии. Если вы проходите курс без оплаты, некоторые задания могут быть недоступны.
Что я получу, оформив подписку на специализацию?
Записавшись на курс, вы получите доступ ко всем курсам в специализации, а также возможность получить сертификат о его прохождении. После успешного прохождения курса на странице ваших достижений появится электронный сертификат. Оттуда его можно распечатать или прикрепить к профилю LinkedIn. Просто ознакомиться с содержанием курса можно бесплатно.
Какие правила возврата средств?
Можно ли получить финансовую помощь?
Learning Outcomes: By the end of this course, you will be able to
Compute the sufficient statistics of a data set that are necessary for learning a PGM from data
Implement both maximum likelihood and Bayesian parameter estimation for Bayesian networks
Implement maximum likelihood and MAP parameter estimation for Markov networks
Formulate a structure learning problem as a combinatorial optimization task over a space of network structure, and evaluate which scoring function is appropriate for a given situation
Utilize PGM inference algorithms in ways that support more effective parameter estimation for PGMs
Implement the Expectation Maximization (EM) algorithm for Bayesian networks
Honors track learners will get hands-on experience in implementing both EM and structure learning for tree-structured networks, and apply them to real-world tasks
Остались вопросы? Посетите Центр поддержки учащихся.
Coursera делает лучшее в мире образование доступным каждому, предлагая онлайн-курсы от ведущих университетов и организаций.
© Coursera Inc., 2019 Все права защищены.
